---
dataset_info:
  features:
  - name: image
    dtype: image
  - name: image_id
    dtype: int64
  - name: answers
    sequence: string
  splits:
  - name: train
    num_bytes: 52294743867.0
    num_examples: 1800000
  download_size: 52155862240
  dataset_size: 52294743867.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---